dumb question about JT-65

[djacobow]

So, I'm pretty new to amateur radio and have been having fun learning the ropes. I've been experimenting with JT65 and I have a question that is a bit provocative, but I hope won't stir up any anger.

Thing is, it really only works if the stations communicating have very accurately -- like sub-second -- synced clocks. Which is all well and good, except that the practical ways to achieve such synchronization (NTP, GPS, whatever) all involved communications with a third party -- in a sense, an out-of-band signal.

So, my question is, isn't that sort of cheating? I mean, is a JT-65 contact really a contact at all?

I understand that lots of contacts benefit from some kind of advance coordination (ie, "the net is on Friday's at 9am" or "I'll be operating from <small_island_nation> from <date> to <date>) but that seems qualitatively different than JT65, which, at least for me, does not work unless I sync my computer's clock pretty much right before use.

[AF6LJ]

When you consider the role the Internet plays in other amateur radio activities such as DX'ing with the DX spotting sites and such, D-Star, C-4 Internet enabled communications and Winlink You could make the argument that JT-65 is much less on the Internet than other modes...
You could do away with the Internet and other external sources at least in theory.

[TheSteve]

I tend to agree there is something not quite right about it. I like to think of radio as being purely radio, no help from the internet, GPS etc.
However that is clearly not the way things are as the internet is such a part of most peoples daily lives so you may as well roll with it.

[djacobow]

When you consider the role the Internet plays in other amateur radio activities such as DX'ing with the DX spotting sites and such, D-Star, C-4 Internet enabled communications and Winlink You could make the argument that JT-65 is much less on the Internet than other modes...
You could do away with the Internet and other external sources at least in theory.

You make a good point.

I sort of wasn't even thinking of the internet-based modes that don't personally rev my engine _because_ the are internet-based. I'm still in the stage where my primary attraction to radio is the somewhat romantic notion of people communicating with nothing in between.

JT-65 does excite me, though, because it's a pretty easy way for an inexperienced op like me to make crazy DX with modest gear and antenna. But that it is tailor-made for exchanging the absolute minimum info for a contact and  this sync issue sort of take a tiny bit of the fun away. Not all the fun, though.

[borjam]

I don't think it's much of a problem. And you don't really need sub second accuracy, sometimes you see people with quite a clock offset

Anyway, if you have a problem with NTP and the Internet (which is legitimate, although it's only depending on the Internet for time synchronization, nothing else) you can also use a GPS receiver for portable operation.

Moreover: if you have a reasonably accurate clock, how many seconds will it drift in several hours? You can set an accurate time, and, once set, disconnect from the Internet and carry on.

Remember that many JT operators run Windows, a toy operating system without a proper NTP implementation, and they must resort to third party solutions. Systems member of the Unix family, such as FreeBSD, Mac OS X or Linux can keep a good synchronization within a few milliseconds without much effort, however.

[Electro Fan]

Does JT65 require both nodes to be physically synchronized with a communications path or just synchronized in time?  (I'm guessing the later.)  If the later, can a JT65 transmitter use a GPSDO or a rubidium clock?  The Internet is less expensive but if you don't want to use the Internet for clock maybe these are alternatives?


http://vi.raptor.ebaydesc.com/ws/eBayISAPI.dll?ViewItemDescV4&item=121966793020&category=73174&pm=1&ds=0&t=1463160441906

http://m.ebay.com/itm/10MHZ-OUTPUT-SINE-WAVE-GPS-DISCiPLINED-CLOCK-GPSDO-GPS-Antenna-Power-supply-/181793953652?nav=SEARCH

http://m.ebay.com/itm/FE-5680A-Rubidium-Atomic-Frequency-Standard-10MHz-SMA-OUTPUT-0-5M-CABLE-/290709958177?nav=SEARCH

[djacobow]

No, JT65 only requires time synchronization, and it's not a super hard requirement. For me, it seems to work much better if my clock is within 1 second of official time, but other have said that even several seconds of slop would still work.

So, yes, you could get a very good clock of any type, set it properly once, and have many lifetimes' of JT65 contacts.

My point was more of an angels-on-pin sort of one: Theoretically, the only way any two clocks can get synchronized is to either use some communications method to sync them, or to bring them together and sync them. The latter is an interesting case. Let's say I have a friend on another continent that I met once on a trip, and I brought along my rubidium clock. We set our clocks together and I went home. Now I want to talk to him over JT65. Voila, no problem -- except for the fact that JT65 conversations are very limited of course. :-)

You could also probably use some mutually observable natural phenomenon as your out-of-band signal. Like the precise moment some distant star went nova. That seems fair game, too. :-)

As I think about it more, it comes down to the concept of "the time" as we understand it. "The time" is not a real thing, but a mutually agreed upon convention that is maintained through a vast and varied network of communications channels. Those channels include, radio, optical, sound ("hey, what time is it?"), physically carrying clocks, etc. But irreducibly, there must be a network for the concept to work.

[babysitter]

Remember that many JT operators run Windows, a toy operating system without a proper NTP implementation, and they must resort to third party solutions. Systems member of the Unix family, such as FreeBSD, Mac OS X or Linux can keep a good synchronization within a few milliseconds without much effort, however.

There are quite proper NTP Implementations for that evil OS, Meinberg has a free one. OTOH, hams which already have 2-way conversion should be able to fix their common clocks to sub-second difference. Good if the donator is close to the rest of the worlds guess what time is.

[Electro Fan]

But irreducibly, there must be a network for the concept to work.

Back in the day I attended lots of meetings with major enterprises that were looking at potential upgrades to large scale networks.  The first hour of almost every new project meeting was spent with the customers and vendors asking "are you going to give clock or are we going to give clock?"  It was like clockwork that this item would come up first regardless of what else was on the agenda.  Until clocking was determined no one seemed very interested in discussing anything else of a technical nature.

[KV6O]

I mean, is a JT-65 contact really a contact at all?

I understand that lots of contacts benefit from some kind of advance coordination (ie, "the net is on Friday's at 9am" or "I'll be operating from <small_island_nation> from <date> to <date>) but that seems qualitatively different than JT65, which, at least for me, does not work unless I sync my computer's clock pretty much right before use.

Sure they count.  Ham's have been using WWV and other time sources world wide for 50+ years to calibrate their dials, really no different than using a NTP to calibrate your computers clock.  And you don't have to do it THAT way, you could have a local rubidium based clock that you would sync ONCE, and it would (in theory) drift no more than a few uS's a MONTH.  Pretty cool!

Steve
KV6O

[AF6LJ]

I mean, is a JT-65 contact really a contact at all?

I understand that lots of contacts benefit from some kind of advance coordination (ie, "the net is on Friday's at 9am" or "I'll be operating from <small_island_nation> from <date> to <date>) but that seems qualitatively different than JT65, which, at least for me, does not work unless I sync my computer's clock pretty much right before use.

Sure they count.  Ham's have been using WWV and other time sources world wide for 50+ years to calibrate their dials, really no different than using a NTP to calibrate your computers clock.  And you don't have to do it THAT way, you could have a local rubidium based clock that you would sync ONCE, and it would (in theory) drift no more than a few uS's a MONTH.  Pretty cool!

Steve
KV6O

Good to see you here Steve.

[KV6O]

I mean, is a JT-65 contact really a contact at all?

I understand that lots of contacts benefit from some kind of advance coordination (ie, "the net is on Friday's at 9am" or "I'll be operating from <small_island_nation> from <date> to <date>) but that seems qualitatively different than JT65, which, at least for me, does not work unless I sync my computer's clock pretty much right before use.

Sure they count.  Ham's have been using WWV and other time sources world wide for 50+ years to calibrate their dials, really no different than using a NTP to calibrate your computers clock.  And you don't have to do it THAT way, you could have a local rubidium based clock that you would sync ONCE, and it would (in theory) drift no more than a few uS's a MONTH.  Pretty cool!

Steve
KV6O

Good to see you here Steve.

You too!  Found my way over here from your post on the Zed! 

Steve
KV6O

[xrunner]

So, my question is, isn't that sort of cheating? I mean, is a JT-65 contact really a contact at all?

Hell yea it's a contact. 

It has nothing to do with the internet and just because your computer is synced to have proper time has no bearing on whether your RF and the remote station RF will make a contact - same as in any other mode. It all depends on the propagation, your power, your antenna ... same as any other mode!

[AF6LJ]

I mean, is a JT-65 contact really a contact at all?

I understand that lots of contacts benefit from some kind of advance coordination (ie, "the net is on Friday's at 9am" or "I'll be operating from <small_island_nation> from <date> to <date>) but that seems qualitatively different than JT65, which, at least for me, does not work unless I sync my computer's clock pretty much right before use.

Sure they count.  Ham's have been using WWV and other time sources world wide for 50+ years to calibrate their dials, really no different than using a NTP to calibrate your computers clock.  And you don't have to do it THAT way, you could have a local rubidium based clock that you would sync ONCE, and it would (in theory) drift no more than a few uS's a MONTH.  Pretty cool!

Steve
KV6O

Good to see you here Steve.

You too!  Found my way over here from your post on the Zed! 

Steve
KV6O

Have a look around; this is a great place and a bunch of really nice folk.
Now you know where I have been hiding out for the last seven months.
There is a hole in the back fence of the Zed's Corn Field.

[djacobow]

So this makes sense:

It has nothing to do with the internet

and this not so much:

and just because your computer is synced to have proper time has no bearing on whether your RF and the remote station RF will make a contact - same as in any other mode. It all depends on the propagation, your power, your antenna ... same as any other mode!

It depends on all those things AND being synced. If you get a contact on JT65 without being synced, it is only because you and your contact are in sync by chance. In that sense, it really is different than other modes because it requires a kind of coordination that other modes do not.

I perhaps stated my question a bit too aggressively. Of course, it's a contact. And frankly, due to the mode's efficiency, it's a way to make some really far-off contacts when doing so in other modes would be hard or impossible.

I accept the point about people tuning up their oscillators using WWV. It is another form of coordination, though at least you do it over the air. Also, you /can/ make contacts even if your dial reads somewhat incorrectly.

The point about being able to set a good clock once and remained synced essentially forver is also a good one.

All that said, though I was initially very excited by JT65, my enthusiasm has waned for this and other reasons. I think other modes are just more fun. That's obviously just an opinion, nothing more.

[tggzzz]

... can keep a good synchronization within a few milliseconds without much effort, however.

And what, exactly, does that mean if the path latency is 200ms? Or if the path latency changes from 200ms to 50ms because the path has changes from a satellite link to an optical fibre link?

Your answer should not violate fundamental theoretical physics. Bonus points for tying it in with Lamport's theorems, see http://amturing.acm.org/award_winners/lamport_1205376.cfm

[borjam]

There are quite proper NTP Implementations for that evil OS, Meinberg has a free one. OTOH, hams which already have 2-way conversion should be able to fix their common clocks to sub-second difference. Good if the donator is close to the rest of the worlds guess what time is.

I am aware of that, but not sure how much proper Windows allows them to be. In order to keep a really good synchronization you need to adjust the clock speed, which UNIX allows with specialized calls. does Windows allow it, or are the ntp implementations just adjusting the time with some frequency?

[xrunner]

It depends on all those things AND being synced. If you get a contact on JT65 without being synced, it is only because you and your contact are in sync by chance.

In that sense, it really is different than other modes because it requires a kind of coordination that other modes do not.

No, you are wrong!

It does not essentially require a kind of coordination that other modes do not. Can you even make a simple SSB contact by not being in sync with the other person's conversation - the other person's comments? No. Try making an SSB contact by responding out of sync or responding 1 minute after the other station says something. It won't work. It's essentially the same problem, it's just that the computer handles it instead of you. So what? The computer handles lots of things these days in many of the modes such as PSK31, RTTY, and others.

I don't know why you are so hung up on the sync part of, like that tends to make it less of a contact. The essential part of radio is still up to propagation and your skill at setting up your station and antenna. And there are tricks to operating that you can use too, it's not that automatic in that way either. It's far from being an automatic contact!

[djacobow]

No, you are wrong!

It does not essentially require a kind of coordination that other modes do not. Can you even make a simple SSB contact by not being in sync with the other person's conversation - the other person's comments? No. Try making an SSB contact by responding out of sync or responding 1 minute after the other station says something. It won't work. It's essentially the same problem, it's just that the computer handles it instead of you. So what? The computer handles lots of things these days in many of the modes such as PSK31, RTTY, and others.

I don't know why you are so hung up on the sync part of, like that tends to make it less of a contact. The essential part of radio is still up to propagation and your skill at setting up your station and antenna. And there are tricks to operating that you can use too, it's not that automatic in that way either. It's far from being an automatic contact!

I think I understand what you saying. Yes, you can sync up to JT65 just by listening. For example, you could set your clock to 0 seconds when you hear the beginning of a transmission. I will try this, but I suspect it works. However, people routinely use JT65 to make contacts with stations that are inaudible, well below the noise floor, so it would not work in all cases. On the other hand, you can use any JT65 transmission to sync up, and they should all start at the same time, so I guess one would be enough to get you going.

Come to think of it, that would be a cool feature to add to the JT65 decoder; the ability to use heard transmissions to adjust a clock used by the app.

I already said it is a contact. But as far as I am concerned, there is something different about this mode because it is literally not just a modulation mode but a time-based communications protocol.